Wildcard week

Group assignment:

Test and automatize the production of nylon thread artificial muscles

Integrate the artificial muscle in fabric


For this week, we were inspired by artificial muscle made of nylon thread. The elegance and simplicity of this solution is very inspiring, and our objective for this week is to make fabable version of this, to demonstrate the possibility of those muscle. More, those artificial muscle can be integrated in fabric, which opens a new world of possibilities!

Some scientific publications:

Making single thread artificial muscles

As reported in the scientific papers, there are multiple ways to include constrains in the thread and make those artificial muscles. We will focus on two ways: even by twisting it on themselves, even by coiling them around an axis.

Solution 1: twisting the thread

Key parameters

To twist the tread, we imagined a small setup controlling the number of turns:

The key parameters are the following:

Basically, we want to know the number of turns to apply and the weight we need to twist the thread. Based on the paper Artificial muscle from fishing line and sewing thread), here is a little script that allows to estimate those values as a function of the thread diameter and its length. Note that this estimation is only valid for small thread diameter (between 0.1mm and 0.3 mm).

After testing it, it overestimates the number of turns for small diameters and overestimates it for large diameter. It can however be used to estimate a number of magnitude.

Thread diameter (mm)
Thread length (mm)
Number of turns
Weight (kg)

Twisting bench

The twisting bench is composed of the following:

The test bench is made modular and designed to be placed on an optical breadboard. The stl files can be downloaded here

code of microcontroller

The thread can directly be fixed on the axis, or through a paperclip, idem for the weight. The setup must be placed on a high enough support (i.e. a table), and the weight must be place to slide along a vertical wall, to limit its rotation. It is possible to set a number of turns using the small interface.

After one or two tests for the calibration, we were able to make easily twisted thread. The thread is subject to break if the weight is too heavy (but the program above is good at this level), or if the number of count is too high! A good solution is to set first a lower number of turns and increase it step by step. After Twisting, the length reduces almost by a factor 4!

Testing the thread

We made some tests to with the fabricated threads. The results were not as impressive as in the papers, but we were able to obtain a contraction of the thread. We suppose that it could be improve by controlling better other parameter (weight, actuation temperature, coil ratio, type of nylon, …).

To obtain contraction, an heavier mass than used for the twisting must be lifted, to elongate a little bit the twisted thread, and allow it to contract. The thread must be first heated once, during which the thread will elongate. The contraction will be obtained when the thread is heated again. We used a hair drier to heat the artificial muscle.

Here is an example of results with a 0.2 mm diameter thread:

With a 0.3 mm diameter thread and a weight of 240g (two times the weight used for twisting):

And with a 0.3 mm diameter thread and a weight of 240g (two times the weight used for twisting). It was this time heated using Joule effect (the thread broke when the heat generation was concentrated). Note that the lower actuation time is due to the low increase of temperature.

Solution 2: coiling around an axis


The other solution to coil to make the artificial muscle is to coil it around an axis. This solution is more suitable for larger thread diameters. We also tested it, and it gave good coiling results!

Testing the thread

However, when testing the thread, we were not able to see contraction. The problem was that no constrained had be included in the coil.

We tried to make coils with constrains, but it broke quite rapidly…

Making a small loom to make fabric

With the working twisted thread, we want to make a small “smart fabric”. To do this, we first made a mini loom frame weaving, found and nicely documented on instructables.

Making the frame with laser cutting:

And making the fabric:

When unloading the fabric from the frame, we have to ensure that the twist in the threads is not released, so we made small laser-cut fixations. If the fabric is not maintained, it will twist macroscopically, due to the constrains of each thread. The laser-cutted files can be downloaded here , or on the instructables page for the frame.

Testing the artificial muscle fabric

We could finally test the fabric. After a first heating causing an elongation, we were able to observe contraction!